The Dewd Robo-Tank build thread

Problems / challenges since last update - These are all with Reef-Pi it seems - and not with Robo-Tank.

I have experienced losing all my temp probes twice in the last two days. Re-loading or re-booting Reef-pi doesn't help. I have to hard power cycle the controller and they come back. A frustration for sure, and potentially a problem, because you lose all your temperature control as well. But if you use the alert function, you can get notified pretty much as soon as it happens and can attend to it.

On my previous controller, I have battled with pH and consistency in readings as well as getting constant strange spikes in readings. This continues with the current controller and I am yet to get to the root of the cause. I know some guys have reported interference from bad / cheap power supplies and replacing them has solved the issue. One guy pinpointed it to down to his lights. His problems started as soon as his lights came on. I thought I had a similar issue with my peristaltic pump on my ATO, but changing how that got power only partly resolved my problems. I did wonder if there was a difference between my previous board and the Robo-Tank pH board. Bottom line is I don't think there is a problem with either, but I do battle with probes holding accurate calibration and I don't actually trust the readings I get, but I do use them to track trends and I have a few things planned to try get better repeatability on pH readings.

Also, I have had a problem with the pH and temp graphs in Reef-Pi. For the last 28 hours or so, the latest time on my pH graph is stuck at 07H00 on 19 January 2021. If I go back on the graph, the latest reading re in an area of the graph about 10% of the way to the left of the latest position. Reef-Pi reports the latest readings correctly to AIO, so my graph there is accurate, but not in Reef-Pi. Yesterday my temps had the same issue, and I just deleted them and re-installed. I'm reluctant to do this with pH as I have spent so much time getting calibration correct. I may still resort to that though.

I used to run a Real Time Clock (RTC) on my old system due to a recommendation. We have fairly regular power cuts here in SA, and apparently it helps with Reef-Pi keeping track of the time when it loses connection to the Internet. It's a simple and cheap addition and I have the board already, so I am considering adding it back to the Robo-Tank installation. And while I'm at it, I have my stand-alone pH board from my old system, so I think I will add it and get an additional pH probe and run two for a while and see if I can pick up any differences and maybe try pin down any pH reading / interference or whatever they might be issues. Will also allow me to move things around and see if that makes any sort of difference to anything. From what I understand I just Daisy-chain them in the I2C connectors.
 
This might be a silly question.....

Can the same float switch be used as the "safety cutoff" for more than one DC Outlet (DC1, 2 or 3 obviously)?

My logic says it should be OK
 
Reason I ask is that my normal "running level ATO" and "Auto Water Change In" use the same sensor so they both fill to the same level. Never on at the same time obviously. Using the same single backup or high level float switch for both outlets makes sense if that can be done.
 
For your temp sensors that could be happening due to a bad sensor, because they share the same bus if one locks it up they all lock up, that would explain them all going down at the same time.

The pH is subject to noise, it is isolated but sometimes noise can still find it's way through. The noise can also be picked up in the air, some light ballasts do this, I recently came across this with someone. When you say it spikes what kind of spike are you talking about? On mine if it happens it can be 0.2 - 0.4 worse case. Calibration sounds like a never ending process, I think I've read it should be done every few months or even a month for best results. Any chance you remember what you put in for point 1 and 2? If so you should be able to just add those number back after deleting the pH.

Not sure about the UI graphs not updating yet the adafruit feeds do, never head of that.

I thought about adding a RTC to controller but opted not to as the modules are easy to add if needed, maybe one day I will.

I don't think you can use the same float on two DC backup ports as it cuts the signal path that controls the mosfet, if float is connected to both when one port switches on I think the other will as well. It would stop them both as expected but I think it would make both signals lines one.
 
There's my pH for the last 24 hours.  Prior to that is a mess because I was testing in buffer solutions and it's all over the place.

And the AIO graphs don't tell the full story, because I've noticed they do more averaging than the reef-pi ones do.


Your comment about sharing a float switch makes sense - kind of why I asked. In my use-case the two outlets would (should!!) never be on at the same time, so it might work / MIGHT never be a problem, but it's easier and safer to just use 2 floats.
 

Attachments

  • 0170b3e5819ca7fabd3774dd6e7a903be9364fe246.jpg
    0170b3e5819ca7fabd3774dd6e7a903be9364fe246.jpg
    670.8 KB · Views: 946
Looks like the 3 big spikes all happened at the same time on the hour, is a pump turning on then? If so where is plugged in and how much current does it draw?

Those other 0.1 spikes will always be there, maybe the frequency can be lowered but due to other things on the board and power supply it'll happen. Looks bad though when zoomed in.

After thinking more about the backup float it won't work on two ports. When ATO pump comes on the signal will cross over to 2nd DC port and turn it on as well.
 
Hmmmmmmm. Good spot there. All the spikes big and small are roughly on the hour. My Kalk stirrer comes on for 5 minutes at the 17th minute of every hour. It's on a AC Outlet, but it's an inductive load. The transformer puts out 24V and after a closer look at it and it looks like a pretty low grade of Chinesium. There is evidence the previous owner "fixed" something on the stirrer. I bought a replacement motor to have as a spare as those are known to have a limited lifespan - guess I'll be changing it out tomorrow. The replacement is a 220V 14W synchronous motor - 60KTYZ One of those 5 rpm motors.

Will report back tomorrow!!!
 
Here's where this stuff gets frustrating for me.  Last night evidence supported the idea that my Kalk Stirrer motor was producing interference that was evident in the pH graph.  Quite definitively I thought.

So I turned that timer off to see if it made a difference.  Smooth graph this morning.  No regular hourly blips.  Proof right?

So I turned the pump on to try reproduce the spike and it won't do it.  Carries on producing the smooth-ish graph.  What I would definitely consider acceptable.

Now it gets interesting.  In "frustration" I turned the Stirrer motor off and suddenly the spike appears in the graph.  My guess is that when turned off, a capacitor or something in the power supply is causing the spike.  Bizarre, but not entirely  inexplicable.

Still going to change the stirrer motor out today and see if that has any effect.
 
That is strange.

I think it can be difficult to trouble shot using charts as they don't update enough, here's something you can do that might help pin point it.

Disable the pH in reef-pi.

Open the Pi terminal window and type in the following commands.

This downloads a python script
sudo wget -c https://robo-tank.ca/Robo-Tank_pH_Script.zip

This unzips the file
sudo unzip Robo-Tank_pH_Script.zip

This runs the script
python Robo-Tank_pH_Script.py

After the last command it’ll say Enter Command: Then type in the following, must be a capital letter.

R

It should print out a pH value. Now type in the following.

POLL,5

Now it should print out the current pH every 5 seconds. Turn things on and and off and see what happens.

Make sure to disable in reef-pi, you can't read at same time from two scripts.
 
Back
Top